With rapid development of a geographic information system, a mobile positioning technology, a wireless communications network, an intelligent terminal technology, and a sensor technology, a situational intelligence application also undergoes rapid development. In the situational intelligence application, how to predict a target geographical position of a mobile user is extremely important for openness of the situational intelligence application.
In the prior art, multiple historical geographical positions that are frequently visited by a mobile user are recorded. When a target geographical position of the mobile user is being predicted, the target geographical position of the mobile user is selected from the multiple historical geographical positions according to a preset prediction model.
However, at an initial stage of prediction, due to a lack of sufficient information about historical geographical positions, selection of the target geographical position of the mobile user may be greatly limited, which leads to a low accuracy of the target geographical position of the mobile user.